Rumor simulation analysis method and system based on large language model agent

Abstract

The invention provides a rumor simulation analysis method and system based on a large language model agent, and the method comprises the steps: determining the feature information of an initialized agent according to a preset propagation network topology structure, and constructing a set number of role agents; based on the to-be-processed speech of the role agent at the current time step, simulating a human thinking activity through a large language model to obtain a variant rumor generated by language propagation of the role agent at the current time step; completing rumor propagation simulation based on the propagation network topology structure and the variant rumors of each time step; in each round of time step, determining the propagation probability between the role agents according to the propagation distance and the credibility of the role agents; and after multiple rounds of iteration, determining the role influence corresponding to each role agent according to the propagation probability and the viewpoint intensity of the role agent. The rumor propagation simulation method and device achieve the technical effect of performing high-precision simulation on the rumor propagation process in the complex social network.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and more specifically, to a rumor simulation analysis method and system based on a large language model intelligent agent. Background Art

[0002] Social networks, as important platforms for information dissemination, play a crucial role in modern society. However, the proliferation of information poses a serious threat to the security and stability of social networks. The spread of rumors on social networks is complex and dynamic, influenced not only by individual cognitive differences and emotional tendencies but also by multiple factors such as social networks and the information environment. Traditional rumor statistical models are significantly inadequate in accurately depicting the dynamic evolution of rumor propagation. Furthermore, existing rumor tracing and intervention technologies mostly rely on retrospective analysis of post-event data and lack the ability to simulate and predict rumor propagation paths in real time.

[0003] Existing technologies exist for real-time rumor simulation and prediction by building rumor propagation simulation models, such as the SIR model and complex network models. However, these models still suffer from the following drawbacks: 1) Behavioral modeling is relatively simplistic, simplifying individuals into homogeneous nodes. However, in actual human language interaction, subjective judgment, emotional expression, and dynamic decision-making play important roles, and traditional models lack high fidelity in rumor simulation scenarios. 2) The network topology is static, making it difficult to adapt to the impact of social network heterogeneity on propagation paths. 3) The interaction mechanism is imperfect, lacking dynamic simulation of complex behaviors such as information correction, secondary creation, and group discussion.

[0004] Therefore, there is an urgent need for a rumor simulation analysis method that can achieve efficient simulation and prediction. Summary of the Invention

[0005] In view of the above problems, the purpose of the present invention is to provide a rumor simulation analysis method and system based on a large language model intelligent agent to solve at least one problem existing in the prior art.

[0006] According to one aspect of the present invention, a rumor simulation analysis method based on a large language model agent is provided, which is applied to electronic devices and includes: According to the preset communication network topology, initialization agent feature information is determined and a set number of role agents are constructed; the initialization agent feature information includes the speech to be processed and the role allocation information of the agent corresponding to each network node; Based on the pending speech of the character agent at the current time step, a large language model is used to simulate human thinking activities to obtain the variant rumors generated by the language propagation of the character agent at the current time step; and rumor propagation simulation is completed based on the topological structure of the propagation network and the variant rumors at each time step. In each time step, the propagation probability between each role agent is determined based on the propagation distance and the credibility of each role agent. After multiple rounds of iteration, the role influence corresponding to each role agent is determined based on the propagation probability and the strength of the role agent's opinion. According to the role influence, key network nodes in the communication network topology are determined.

[0007] In addition, an optional technical solution is that the roles include ordinary users, opinion leaders and rumor makers.

[0008] In addition, an optional technical solution is that the topology of the propagation network includes one or more of a small-world network, a scale-free network, and a random network; When the number of the propagation network topology structures is two or more, the propagation network topology structures are dynamically switched.

[0009] In addition, the optional technical solution is, Based on the pending speech of the character agent at the current time step, the variant rumors generated by the language propagation of the character agent at the current time step are obtained by simulating human thinking activities through a large language model This is achieved through the following formula:

[0010] Among them, the large language model is the LLM large language model, For the role agent, is a mutated rumor; the basic rumor is the pending speech of the role agent at the current time step, and the context information includes historical propagation records and node interaction behaviors; the information of the role agent includes age, gender, and work status.

[0011] In addition, an optional technical solution is to determine the propagation probability between each role agent based on the propagation distance and the credibility of each role agent, which can be achieved through the following formula:

[0012] in, For the network The role agent in For nodes Neighbors; For character agents credibility; is the shortest path distance between nodes, is the spatial attenuation coefficient.

[0013] In addition, an optional technical solution is to realize the credibility of each character agent through the following formula:

[0014] in, is the role influence coefficient, For the network The intelligent agent in .

[0015] In addition, an optional technical solution is that the opinion strength of the role agent is realized by the following formula:

[0016] Among them, when the role agent To the neighbors When a rumor is successfully spread, The strength of opinion is expressed as , strength of opinion ∈[0,1] represents the subjective acceptance of the agent to the rumor; For character agents credibility, is the role acceptance weight.

[0017] On the other hand, the present invention also provides a rumor simulation analysis system based on a large language model agent, which performs rumor simulation analysis using the rumor simulation analysis method based on a large language model agent as described above; the system comprises: A role agent construction unit is used to determine initial agent feature information and construct a set number of role agents based on a preset communication network topology; the initial agent feature information includes the speech to be processed and the role allocation information of the agent corresponding to each network node; A propagation simulation unit is configured to simulate human thinking activities using a large language model based on the pending speech of the role agent at the current time step to obtain variant rumors generated by the language propagation of the role agent at the current time step; and complete rumor propagation simulation based on the topological structure of the propagation network and the variant rumors at each time step; The role influence determination unit is used to determine the propagation probability between each role agent according to the propagation distance and the credibility of each role agent in each time step; after multiple rounds of iterations, the role influence corresponding to each role agent is determined according to the propagation probability and the opinion strength of the role agent; based on the role influence, the key network nodes in the topological structure of the propagation network are determined.

[0018] The rumor simulation analysis method and system based on the large language model intelligent agent of the present invention fully considers factors such as subjective judgment, emotional expression, and dynamic decision-making in the process of human language interaction in terms of behavioral modeling, and successfully overcomes the limitations of traditional models that simplify individuals into homogeneous nodes, making the behavior of the intelligent agent closer to the performance of real humans in rumor propagation. It supports different network topologies such as small-world networks, scale-free networks, and random networks, and quantifies the influence of different users on the spread of rumors through role allocation and credibility calculation. It achieves the technical effect of high-precision simulation of the rumor propagation process in complex social networks and provides a scientific basis for rumor governance.

[0019] In order to achieve the above and related purposes, one or more aspects of the present invention include the features that will be described in detail later. The following description and the accompanying drawings describe some exemplary aspects of the present invention in detail. However, these aspects indicate only some of the various ways in which the principles of the present invention can be used. In addition, the present invention is intended to include all of these aspects and their equivalents. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] By referring to the following description in conjunction with the accompanying drawings, and with a more complete understanding of the present invention, other objects and results of the present invention will become more clear and easy to understand. In the accompanying drawings: Figure 1 2. A flowchart of a rumor simulation method based on a large language model agent according to an embodiment of the present invention; Figure 2 A node degree distribution diagram of a network structure of a rumor simulation analysis method based on a large language model agent according to an embodiment of the present invention; Figure 3 This is an influence analysis diagram of different roles in a rumor simulation analysis method based on a large language model agent according to an embodiment of the present invention; Figure 4 A schematic diagram of a module of a rumor simulation and analysis system based on a large language model agent provided by one embodiment of the present invention; Figure 5 A schematic diagram of the internal structure of an electronic device that implements a rumor simulation analysis method based on a large language model agent, provided by one embodiment of the present invention.

[0021] The same reference numerals throughout the drawings indicate similar or corresponding features or functions. DETAILED DESCRIPTION

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0023] The following is a clear and detailed description of the technical solutions in the embodiments of the present application, with reference to the accompanying drawings. In the description of the embodiments of the present application, unless otherwise specified, the term "and / or" is merely a description of the association relationship between associated objects, indicating that three relationships can exist. For example, "A and / or B" can represent the following three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0024] In the following, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of the technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of such features. Furthermore, in the description of the embodiments of this application, "plurality" means two or more than two.

[0025] References to "one embodiment" or "some embodiments" in this specification mean that a particular feature, structure, or characteristic described in conjunction with that embodiment is included in one or more embodiments of the present application. Thus, phrases such as "in one embodiment," "in some embodiments," "in other embodiments," and "in yet other embodiments" appearing in various places in this specification do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized. The terms "including," "comprising," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0026] This paper proposes a multi-agent simulation framework based on a large language model (LLM). Through the LLM, each agent is endowed with integrated "cognition, emotion, and behavior" decision-making capabilities, enabling it to generate rumor content consistent with human language logic (such as emotional expressions and logical loopholes) and simulate individuals' dynamic behaviors during the dissemination process, such as questioning, forwarding, and correction, effectively enhancing the realism of the simulation scenario. The system can dynamically switch between different structures, such as small-world networks, scale-free networks, and random networks, to quantitatively analyze the impact of network heterogeneity on rumor propagation efficiency, coverage, and lifecycle. Furthermore, this paper supports dynamic interaction and evolutionary analysis. Leveraging real-time interaction between agents, it tracks the entire process of rumor generation, mutation, and extinction, extracting key dissemination nodes (such as "opinion leaders"), information evolution paths, and group polarization phenomena, providing data support for the development of precise rumor-debunking strategies. Compared with traditional methods, this paper combines LLM with agents, achieving a paradigm shift from "static statistical prediction" to "dynamic behavioral simulation," providing multi-dimensional technical support for public opinion monitoring, network governance policy design, and social science research.

[0027] In order to describe in detail the rumor simulation analysis method and system based on a large language model intelligent agent of the present invention, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Example 1

[0028] Figure 1 A flowchart of a rumor simulation analysis method based on a large language model agent according to an embodiment of the present invention is shown.

[0029] like Figure 1 As shown, the rumor simulation analysis method based on a large language model agent according to an embodiment of the present invention includes steps S110 to S130.

[0030] S110: According to the preset communication network topology, determine the initialization agent feature information and construct a set number of role agents; wherein the initialization agent feature information includes the speech to be processed and the role allocation information of the agent corresponding to each network node.

[0031] When encountering a speech dissemination scenario, a multi-agent large language model can simulate and analyze the speech (rumor) propagation process of the event, providing early warning of the risk of rumor spread on social networks. First, the topological network must be initialized and the characteristics of each agent in the multi-agent network must be set. This topological network depicts the social relationships between agents, with each node representing an agent, and each agent simulating the behaviors and characteristics of a specific role. The role assignment information for agents can include not only settings for ordinary users, opinion leaders, rumor mongers, and more, but also information such as age, gender, and employment status.

[0032] In a specific embodiment, the propagation network topology includes one or more of a small-world network, a scale-free network, and a random network; when the number of the propagation network topology is more than two, the propagation network topology is dynamically switched. In terms of the heterogeneity of social networks, the present invention abandons the traditional static network topology and adopts a dynamically adaptive network topology model, which can accurately capture the impact of social network heterogeneity on the rumor propagation path, thereby more accurately reflecting the propagation law of rumors in different social network environments. In the specific implementation process, different network structures, such as small-world networks, scale-free networks, etc., are generated through toolkits such as networkx, and then evolution is carried out on this network. At the same time, the nodes in the network are numbered and correspond to the intelligent agents of the large language model. Therefore, it is possible to set the type of evolved network structure. That is, the intelligent agents are connected to each other through a preset network topology, and support the number of neighbors. k and reconnection probability β Small-world networks control local clustering and global connectivity, generating a "hub-node" structure through a "preferential connection" mechanism. This simulates the heterogeneous scale-free networks of social platforms and random networks that randomly establish connections between nodes. For example, in a small-world network, a node is set to spread a rumor, which then spreads to neighboring nodes according to certain rules. The diffusion time is fixed, for example, 20 times, which is 20 time steps. Each time, the number of people infected by the rumor on the network is counted.

[0033] Each agent is assigned a specific role, including "ordinary user", "opinion leader", "rumor maker", etc. Among them, some roles are randomly assigned through probability distribution (such as ordinary users accounting for 80%), and high-connectivity nodes in the network position (such as hub nodes in scale-free networks) are preferentially assigned to the "opinion leader" role to simulate the characteristics of influential nodes in social networks. In addition, credibility modeling information is added when constructing role agents, and the social network attributes of the node are combined with the role weight to calculate the credibility. Specifically, credibility is a quantitative indicator of its influence in spreading rumors, which is determined by the node network attributes and role weight. The network attribute of the node is the node's degree, which represents the influence of the node in the current environment, and the role is the role setting of the node. For example, in the traffic accident rumor propagation scenario, the police role credibility is 0.8.

[0034] The credibility of each character agent is achieved through the following formula:

[0035] in, is the role influence coefficient (e.g. opinion leader value is 2, ordinary user value is 1), For the network Ensure that nodes with high connectivity or high role weight have higher credibility, u For the network V The intelligent agent in .

[0036] In general, the present invention is to build a heterogeneous multi-agent system based on a large language model, and realize a high-fidelity simulation of the rumor propagation process through agent role allocation, credibility modeling, opinion strength evolution and network connection mechanism. In terms of improving the interactive mechanism, the present invention innovatively introduces the dynamic simulation of complex behaviors such as information correction, secondary creation and group discussion, which greatly enriches the interactive details of the rumor propagation simulation, so that the simulation process can more comprehensively and realistically reflect the various complex phenomena in the spread of rumors. By integrating the semantic generation capabilities of large language models with the collaborative simulation advantages of multi-agent systems, the present invention breaks through the limitations of traditional technologies in behavioral simulation, network adaptation and dynamic interaction, and provides innovative technical means for the research and governance of rumor propagation in social networks.

[0037] S120: Based on the pending speech of the role agent at the current time step, the mutated rumors generated by the language propagation of the role agent at the current time step are obtained by simulating human thinking activities through a large language model; and the rumor propagation simulation is completed based on the topological structure of the propagation network and the mutated rumors at each time step.

[0038] Based on the pending speech of the role agent at the current time step, the variant rumor generated by the language propagation of the role agent at the current time step obtained by simulating human thinking activities through a large language model is realized by the following formula:

[0039] Among them, the large language model is the LLM large language model, For the role agent, is a mutated rumor; the basic rumor is the pending speech of the role agent at the current time step; the context information includes historical propagation records and node interaction behaviors; the information of the role agent includes age, gender, and work status.

[0040] It should be noted that a Large Language Model (LLM) is a deep learning model based on artificial neural networks, primarily used for natural language processing (NLP) tasks. LLMs are typically based on the Transformer architecture, whose core feature is the self-attention mechanism. This allows the model to dynamically focus on relationships between different positions in a sequence when processing sequential data, thereby capturing long-range dependencies. LLMs typically consist of multiple layers of Transformer encoders or decoders, each with multiple attention heads (multi-head attention), which can focus on different feature dimensions in parallel. The model size can be increased by increasing the number of layers, the number of attention heads, and the number of neurons in each layer. LLMs are trained through unsupervised learning on large text corpora. The training objective is typically language modeling, which involves predicting the next word or character in a text sequence given the previous words or characters. In this way, the model can learn the statistical regularities, semantic information, and grammatical structure of the language.

[0041] In this invention, a large language model (LLM) is used to simulate human thought processes based on the speech of a character agent, thereby selecting and generating variant rumors through language propagation. To fine-tune the large language model (LLM) based on the application scenarios of this invention and better adapt it to the specific needs of rumor propagation simulation, the following steps can be followed to generate realistic rumor content and simulate agent behavior. First, data collection and preprocessing are performed. Specifically, public opinion data, including user basic information, event participation, comments, reposts, and likes, is collected from platforms such as social media and news websites, and the data is cleaned and preprocessed. Next, training data is prepared and training samples are constructed. The input consists of user personal information, event participation, previous comments, and behavior, and the output consists of actual user comments, reposts, or likes for specific events. The samples are annotated. Finally, fine-tuning objectives are defined. Specifically, these objectives include generating rumor content that conforms to human language logic and simulating agent propagation behavior. A pretrained LLM model, such as GPT-3, GPT-4, or BERT, is selected as the base model, and the model parameters are adjusted according to the task requirements. After initializing the model, fine-tune it using the prepared training data. Using optimization algorithms (such as LoRA and PPO), the model parameters are updated to minimize the loss function. Appropriate evaluation metrics, such as perplexity and BLEU scores, are selected, and the model is validated using a validation set. Finally, the fine-tuned model is deployed in the simulation system and regularly updated based on feedback and new data. This fine-tuned LLM can more accurately simulate the behavior and speech generation process of intelligent agents in rumor spreading, improving the reliability and realism of the simulation system.

[0042] S130: In each time step, the probability of communication between the various role agents is determined based on the communication distance and the credibility of each role agent. After multiple rounds of iteration, the role influence corresponding to each role agent is determined based on the communication probability and the strength of the role agent's opinion.

[0043] The implementation process of the opinion strength evolution mechanism first sets the initial value of opinion strength, which is set by the communicator (for example, the initial communicator value is 0.8); then, in subsequent propagation, due to propagation attenuation, the opinion strength of each role agent generated by propagation attenuation in subsequent propagation is realized by the following formula:

[0044] Among them, when the role agent To the neighbors When a rumor is successfully spread, The strength of opinion is expressed as , strength of opinion ∈[0,1] represents the subjective acceptance of the agent to the rumor; For character agents credibility, is the role acceptance weight. For example, the role acceptance weight of the opinion leader value The role acceptance weight of 0.9 is the value of the average user It is 0.4 to reflect the different sensitivities of different characters to rumors.

[0045] The present invention sets up an intelligent agent propagation trigger mechanism for rumor spreading. That is, in each round of time step, the infected nodes Try asking your neighbors Rumor propagation. The agent-based triggering mechanism for rumor propagation embodies a "neighbor-first" approach, meaning that the greater the distance or the less credible the propagator, the lower the probability of propagation. Based on this propagation probability, combined with the propagation intensity, a threshold is set to control whether a node mutates.

[0046] The following formula is used to determine the propagation probability between each role agent based on the propagation distance and the credibility of each role agent:

[0047] in, For the network The role agent in For nodes Neighbors; For character agents credibility; is the shortest path distance between nodes, is the spatial attenuation coefficient.

[0048] The pseudo code of the rumor simulation analysis method based on the large language model agent of the present invention is shown in Table 1: Table 1 Process pseudo code

[0049]

[0050] The present invention's rumor simulation and analysis method based on a large language model agent is designed to construct a heterogeneous multi-agent system based on a large language model. Through agent role assignment, credibility modeling, opinion strength evolution, and network connection mechanisms, it achieves a highly realistic simulation of the rumor propagation process. By integrating the semantic generation capabilities of a large language model with the collaborative simulation advantages of a multi-agent system, the present invention overcomes the limitations of traditional technologies in behavioral simulation, network adaptation, and dynamic interaction, providing innovative technical means for the research and governance of rumor propagation on social networks. Example 2

[0051] In this example, a fixed initial rumor topic, "A serious car accident occurred at a certain intersection," serves as the starting point for all rumor propagation. Subsequent spreaders will generate variants with character traits based on this topic. First, agent roles are assigned. Ten roles are defined, covering different social identities, and each role is assigned a credibility value. A higher value indicates greater influence. The specific roles and their influences are as follows: "Pedestrian": 0.3, "Eyewitness": 0.8, "Traffic Police": 0.9, "Medical Staff": 0.9, "Driver of the Accident Vehicle": 0.6, "Nearby Shopkeeper": 0.4, "Reporter": 0.7, "Nearby Resident": 0.4, "Rescue Worker": 0.8, and "Taxi Driver": 0.5. Traffic police, medical staff, and other roles are considered high-credibility roles, while passersby and other roles are considered low-credibility roles. It should be noted that in the specific implementation, each node in the network corresponds to a role for each role. The specific values for the role distribution can be customized, for example, setting 2 police officers and 20 passersby.

[0052] Next, we set the simulation parameters. Specifically, we used the language model API provided by Zhipu, selected the GLM-4-FLASH language model, set the generation temperature to 0.9, and used a small-world network with an edge reconnection probability of 0.1, a scale-free network with 2 edges per new node, and a random network with an edge existence probability of 0.1. We also set the agent format to 100 and the number of time steps to 10 for the simulation.

[0053] Then, the network topology is dynamically adapted. In this embodiment, the propagation network topology is dynamically switched among the small-world network, scale-free network, and random network. The node degree distribution of the three propagation network topologies is shown in the following figure. Figure 2 shown.

[0054] By observation Figure 2 As can be seen, small-world networks exhibit an "inverted U-shaped" distribution, with the probability approaching 0 in low-degree regions (k < 5), reaching a peak at k ≈ 15 (P ≈ 0.12), and then rapidly decaying. Scale-free networks exhibit the rudiments of a power-law distribution, exhibiting linear decay in the range k = 10^0 to 10^1. Random networks are perfect bell-shaped curves with a peak k of 20. A variety of network settings can demonstrate the robustness of this simulation model. Regarding the heterogeneity of social networks, this paper abandons the traditional static network topology and adopts a dynamically adaptive network topology model. This model can accurately capture the impact of social network heterogeneity on rumor propagation paths, thereby more accurately reflecting the spread of rumors in different social network environments.

[0055] Finally, in the rumor propagation scenario, role influence analysis is performed. It should be noted that the higher the average opinion strength, the more likely the role is to become an "information amplifier" in the propagation chain (e.g., the traffic police have high initial strength and high credibility). The spread of rumors is more likely to be accepted by subsequent nodes (e.g., the strength of medical staff decays slowly). After 10 rounds of simulation in this embodiment, the influence distribution of different roles represented by each agent is as follows: Figure 3 shown.

[0056] By observation Figure 3 As can be seen, journalists, nearby residents, and nearby shop owners have a higher influence. This is because these three groups are more likely to have visited the accident scene and have a high presence online, making them a stronger influence on public opinion. Pedestrians and taxi drivers, on the other hand, have relatively low influence due to their initial credibility. While traffic police, medical staff, and other groups have higher initial credibility, their professions make them more objective and rational in their statements during the simulation, making them less influential on the spread of rumors.

[0057] like Figure 4 As shown, the present invention provides a rumor simulation analysis system based on a large language model agent, which uses the rumor simulation analysis method based on a large language model agent as described above to perform rumor simulation analysis based on a large language model agent. According to the functions implemented, the rumor simulation analysis system 400 based on a large language model agent may include a role agent construction unit 410, a propagation simulation unit 420, and a role influence determination unit 430. The unit of the present invention can also be called a module, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0058] In this embodiment, the functions of each module / unit are as follows: The role agent construction unit 410 is used to determine the initialization agent feature information and construct a set number of role agents based on the preset communication network topology; the initialization agent feature information includes the speech to be processed and the role allocation information of the agent corresponding to each network node.

[0059] The communication simulation unit 420 is used to simulate human thinking activities through a large language model based on the pending speech of the role agent at the current time step to obtain the mutated rumors generated by the language communication of the role agent at the current time step; and complete the rumor propagation simulation based on the topological structure of the communication network and the mutated rumors at each time step.

[0060] The role influence determination unit 430 is used to determine the propagation probability between each role agent according to the propagation distance and the credibility of each role agent in each time step; after multiple rounds of iterations, the role influence corresponding to each role agent is determined according to the propagation probability and the opinion strength of the role agent; based on the role influence, the key network nodes in the topological structure of the propagation network are determined.

[0061] The rumor simulation and analysis system based on the large language model intelligent agent of the present invention fully considers factors such as subjective judgment, emotional expression, and dynamic decision-making in the process of human language interaction in terms of behavioral modeling, and successfully overcomes the limitations of traditional models that simplify individuals into homogeneous nodes, making the behavior of the intelligent agent closer to the performance of real humans in rumor propagation. It supports different network topologies such as small-world networks, scale-free networks, and random networks, and quantifies the influence of different users on the spread of rumors through role allocation and credibility calculation. It achieves the technical effect of being able to simulate the rumor propagation process in complex social networks with high precision and provide a scientific basis for rumor governance.

[0062] For more specific implementation methods of the rumor simulation analysis system based on a large language model agent, please refer to the aforementioned description of the embodiment of the rumor simulation analysis method based on a large language model agent, and will not be described in detail here.

[0063] like Figure 5 As shown, the present invention also provides an electronic device 1 for a rumor simulation analysis method based on a large language model intelligent agent.

[0064] The electronic device 1 may include a processor 10, a memory 11, and a bus. It may also include a computer program stored in the memory 11 and executable on the processor 10, such as a rumor simulation analysis program 12 based on a large language model agent. The memory 11 may also include both an internal storage unit of the rumor simulation analysis system based on a large language model agent and an external storage device. The memory 11 may be used not only to store installed application software and various types of data, such as the code for the rumor simulation analysis program based on a large language model agent, but also to temporarily store data that has been output or is about to be output.

[0065] The memory 11 includes at least one type of readable storage medium, including flash memory, a removable hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory 11 may be an internal storage unit of the electronic device 1, such as a removable hard disk of the electronic device 1. In other embodiments, the memory 11 may also be an external storage device of the electronic device 1, such as a plug-in removable hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a flash memory card, etc. equipped on the electronic device 1. Furthermore, the memory 11 may include both an internal storage unit of the electronic device 1 and an external storage device. The memory 11 can be used not only to store application software installed on the electronic device 1 and various types of data, such as the code of a rumor simulation analysis program based on a large language model agent, but also to temporarily store data that has been output or is about to be output.

[0066] In some embodiments, the processor 10 may be comprised of an integrated circuit, such as a single packaged integrated circuit or multiple packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and a combination of various control chips. The processor 10 is the control core (control unit) of the electronic device, connecting the various components of the electronic device using various interfaces and circuits. It executes programs or modules stored in the memory 11 (e.g., a rumor simulation analysis program based on a large language model agent) and accesses data stored in the memory 11 to perform various functions and process data.

[0067] The bus may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is configured to enable connection and communication between the memory 11 and at least one processor 10, etc.

[0068] Figure 5 Only the electronic device with components is shown, and it can be understood by those skilled in the art that Figure 5The structure shown does not constitute a limitation on the electronic device 1 , and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0069] For example, although not shown, the electronic device 1 may further include a power source (e.g., a battery) to power various components. Preferably, the power source may be logically connected to the at least one processor 10 via a power management system, thereby enabling functions such as charge management, discharge management, and power consumption management through the power management system. The power source may further include any components such as one or more DC or AC power sources, a recharging system, a power failure detection circuit, a power converter or inverter, and a power status indicator. The electronic device 1 may further include various sensors, Bluetooth modules, Wi-Fi modules, etc., which are not further described here.

[0070] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device 1 and other electronic devices.

[0071] Optionally, the electronic device 1 may further include a user interface, which may be a display or an input unit (such as a keyboard). Optionally, the user interface may also be a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display may also be appropriately referred to as a display screen or a display unit, and is used to display information processed by the electronic device 1 and to display a visual user interface.

[0072] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0073] The rumor simulation analysis program 12 based on the large language model agent stored in the memory 11 of the electronic device 1 is a combination of multiple instructions. When running in the processor 10, it can achieve: according to the preset propagation network topology, determine the initialization agent feature information and construct a set number of role agents; the initialization agent feature information includes the speech to be processed and the role allocation information of the agent corresponding to each network node; based on the speech to be processed of the role agent at the current time step, simulate the human thinking activity through the large language model to obtain the variant rumor generated by the language propagation of the role agent at the current time step; complete the rumor propagation simulation based on the propagation network topology and the variant rumor of each time step; in each round of time step, determine the propagation probability between each role agent according to the propagation distance and the credibility of each role agent; after multiple rounds of iterations, determine the role influence corresponding to each role agent according to the propagation probability and the opinion strength of the role agent; and determine the key network nodes in the propagation network topology according to the role influence.

[0074] Specifically, the specific implementation method of the processor 10 for the above instructions can refer to Figure 1 The description of the relevant steps in the corresponding embodiments is not repeated here. Furthermore, if the modules / units integrated in the electronic device 1 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable medium may include: any entity or system capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM).

[0075] An embodiment of the present invention also provides a computer-readable storage medium, which can be non-volatile or volatile, and stores a computer program. When the computer program is executed by a processor, it implements: according to a preset communication network topology, determining the initialization agent feature information and constructing a set number of role agents; wherein, the initialization agent feature information includes the speech to be processed and the role allocation information of the agent corresponding to each network node; based on the speech to be processed of the role agent at the current time step, simulating human thinking activities through a large language model to obtain the variant rumors generated by the language communication of the role agent at the current time step; completing the rumor propagation simulation based on the communication network topology and the variant rumors at each time step; in each round of time step, determining the propagation probability between each role agent based on the propagation distance and the credibility of each role agent; after multiple rounds of iterations, determining the role influence corresponding to each role agent based on the propagation probability and the strength of the opinion of the role agent; and determining the key network nodes in the communication network topology based on the role influence.

[0076] Specifically, the specific implementation method when the computer program is executed by the processor can refer to the description of the relevant steps in the rumor simulation analysis method based on the large language model intelligent agent in the embodiment, which will not be repeated here.

[0077] In the several embodiments provided herein, it should be understood that the disclosed devices, systems, and methods may be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the module division is merely a logical functional division, and actual implementation may employ other division methods.

[0078] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed across multiple network elements. Some or all of the modules may be selected to achieve the purpose of the solution of this embodiment according to actual needs.

[0079] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or hardware plus software functional modules.

[0080] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0081] Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the claims are intended to be embraced therein. Any reference to a figure in a claim should not be construed as limiting the claim to which it relates.

[0082] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units or systems stated in the system claim can also be implemented by one unit or system through software or hardware.

[0083] However, those skilled in the art will appreciate that various improvements can be made to the rumor simulation analysis method and system based on a large language model agent, without departing from the scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the contents of the appended claims.

Claims

1. A rumor simulation analysis method based on a large language model agent, applied to electronic devices, characterized in that: include: According to the preset communication network topology, determine the initialization agent feature information and construct a set number of role agents; wherein the initialization agent feature information includes the speech to be processed and the role allocation information of the agent corresponding to each network node; Based on the pending speech of the character agent at the current time step, a large language model is used to simulate human thinking activities to obtain the variant rumors generated by the language propagation of the character agent at the current time step; and rumor propagation simulation is completed based on the topological structure of the propagation network and the variant rumors at each time step. In each time step, the propagation probability between each role agent is determined based on the propagation distance and the credibility of each role agent; after multiple rounds of iterations, the role influence corresponding to each role agent is determined based on the propagation probability and the opinion strength of the role agent.

2. The rumor simulation analysis method based on a large language model agent according to claim 1 is characterized in that: The roles include ordinary users, opinion leaders and rumor makers.

3. The rumor simulation analysis method based on a large language model agent according to claim 1 is characterized in that: The propagation network topology includes one or more of a small-world network, a scale-free network, and a random network; When the number of the propagation network topology structures is two or more, the propagation network topology structures are dynamically switched.

4. The rumor simulation analysis method based on a large language model agent according to claim 1 is characterized in that: Based on the pending speech of the role agent at the current time step, the variant rumor generated by the language propagation of the role agent at the current time step is obtained by simulating human thinking activities through a large language model, which is achieved through the following formula: ， Among them, the large language model is the LLM large language model, For the role agent, is a mutated rumor; the basic rumor is the pending speech of the role agent at the current time step, and the context information includes historical propagation records and node interaction behaviors; the information of the role agent includes age, gender, and work status.

5. The rumor simulation analysis method based on a large language model agent according to claim 1 is characterized in that: The propagation probability between each role agent is determined based on the propagation distance and the credibility of each role agent, which is achieved through the following formula: ， in, For the network The role agent in For nodes Neighbors; For character agents credibility; is the shortest path distance between nodes, is the spatial attenuation coefficient.

6. The rumor simulation analysis method based on a large language model agent according to claim 5 is characterized in that: The credibility of each character agent is achieved through the following formula: ， in, is the role influence coefficient, For the network The intelligent agent in .

7. The rumor simulation analysis method based on a large language model agent according to claim 1 is characterized in that: The opinion strength of the character agent is realized by the following formula, ， Among them, when the role agent To the neighbors When a rumor is successfully spread, The strength of opinion is expressed as , strength of opinion ∈[0,1] represents the subjective acceptance of the agent to the rumor; For character agents credibility, is the role acceptance weight.

8. A rumor simulation analysis system based on a large language model agent, using the rumor simulation analysis method based on a large language model agent as described in any one of claims 1 to 7 to perform rumor simulation analysis; comprising: A role agent construction unit is used to determine initial agent feature information and construct a set number of role agents based on a preset communication network topology; the initial agent feature information includes the speech to be processed and the role allocation information of the agent corresponding to each network node; A propagation simulation unit is configured to simulate human thinking activities using a large language model based on the pending speech of the role agent at the current time step to obtain variant rumors generated by the language propagation of the role agent at the current time step; and complete rumor propagation simulation based on the topological structure of the propagation network and the variant rumors at each time step; The role influence determination unit is used to determine the propagation probability between each role agent based on the propagation distance and the credibility of each role agent in each time step; After multiple rounds of iterations, the role influence corresponding to each role agent is determined based on the propagation probability and the opinion strength of the role agent; According to the role influence, key network nodes in the communication network topology are determined.

9. An electronic device, characterized in that: The electronic device includes a memory, a processor, and a rumor simulation analysis program based on a large language model agent stored in the memory and runnable on the processor. When the rumor simulation analysis program based on a large language model agent is executed by the processor, the rumor simulation analysis method based on a large language model agent as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the rumor simulation analysis method based on a large language model agent as described in any one of claims 1 to 7 is implemented.

Images